Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Analysis of barley by NIRS

item Himmelsbach, David
item Sohn, Mi Ryeong
item Hicks, Kevin
item Barton Ii, Franklin

Submitted to: UJNR Food & Agricultural Panel Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 9/28/2006
Publication Date: 10/18/2006
Citation: Himmelsbach, D.S., Sohn, M., Hicks, K.B., Barton II, F.E. 2006. Analysis of barley by NIRS. United States-Japan Natural Resources Food & Agricultural Panel; October 21-28, Sonoma,CA. Pp IA1-IA5.

Interpretive Summary: Barley is being considered as alternative grain to corn for the production of bioethanol. A rapid method of analysis of barley constituents is needed to evaluate its use as a raw material. Unlike corn barley that contains starch as the major carbohydrate source, barley also a considerable amount of Beta-glucan. This generates an additional time consuming assay along with others required. Rapid overall analyses by near-infrared spectroscopy (NIRS) have been developed that eliminate the need for long assays. Good correlation to Beta-glucan on whole grains required a high-resolution spectrometer whereas a low-resolution spectrometer sufficed for the analysis of: moisture, starch, protein, oil and ash. The NIRS analysis can be accomplished in about a minute.

Technical Abstract: Development of a rapid method of analysis of barley for moisture, starch, protein, oil, ash and Beta-glucan was attempted. One hundred forty-three barley grain samples of 3 types (hulled, hulless and malt) over 2 growing seasons and from various locations in the United States were utilized in the study. Both whole and ground samples were investigated using three different near-infrared spectroscopy (NIRS) instruments. Samples could be classified by principal component analysis (PCA) by type and growing season. Good correlations (R2>0.9) were obtained to moisture, starch, protein and oil compositional data with lower correlations (R2~0.8) to ash and Beta-glucan using multivariate partial least squares-1 (PLS1). Results on ground samples were better than for whole grain samples. The highest resolution instrument produced the best prediction results for all components but oil and ash, which seem to require the inclusion of the visible region of the spectrum.

Last Modified: 10/15/2017
Footer Content Back to Top of Page